[Sorry for the delayed and necessarily very partial response - can't do more right now. ]
1) The press stories for the 2004 SCF data are based only on a summary report (http://www.federalreserve.gov/pubs/bulletin/2006/financesurvey.pdf) -- in fact they seem based on the press handouts produced by the Fed's media department. The full public data for the SCF won't be released until March 3rd (and most progressive researchers need the computer tapes themselves to overcome some of the items left out of the public version).
From what I have seen from Google News the press reports reflect the Fed's press kits' emphasis on the national average or median (i.e. little progress for the typical family) -- not the additional and different picture when one looks separately at the rich and everyone else.
2) One SCF specialty is household wealth holdings (home value, stocks, bonds). It is true that 2001-04 were not the go-go years, but typical households in the top 10 percent of income (making above $129k p/y) increased their household wealth by over $36 thousand during the 3 years. In stark contrast, 60% of the country (making under $53k p/y) enjoyed no such progress over 3 years of a "healthy" economy... they actually lost a bit of wealth!
[Based on previous SCF releases, a more detailed analysis of the 2004 SCF will likely show that these numbers mask an even greater decline in equality. The SCF public numbers (unlike the full Survey) don't seem to include the massive loss of pension wealth by working class families due to the decline in defined benefit pension plans. Any coming housing value deflation may also take a disproportionate toll here. And the wealth of the very rich is somewhat excluded from this survey. ]
The pattern of declining equality is also borne out by the new annual income figures. Here too the top 10% did well (although not like in the '90s) - their median income grew a hefty $4,200 per year. But it looks like the remaining 90% of the country (making under $129k p/y) taken together saw their median annual income drop (simply averaging the quintile medians).
[There are variations within quintile groups and, as Doug Henwood points out, one quintile group (the 40-60% group, making between $34k and $53k p/y) had a small $700 increase over the 3 years. That sole slight anomaly might be clearer when the data is available. As Doug also points out, the lowest quintile (making under $19k p/y) also had a slight rise of $200. But this has been going on for 10-15 years and previously was shown to be in good part a statistical artifact: some better off retirees appear in this bracket because of their low current earnings and they have benefited from higher property income. In previous SCF Surveys, the others in this bracket showed a decline in income (one could also add that Soc Sec wasn't cut).]
3) C.G. Eastbrook asks about the latest SCF and trends during Bush's time in office. I have observed (with no great expertise) that in the early days of neo-liberalism (say during the Reagan Years) income/wealth shifted first to the top 40%. Then, as the new arrangements consolidated (say during the GWB-Clinton years), mainly the top 20% benefited. Lately, the shift may have been limited to the top 10% (or even top 1-3%). It *sounds* like the latest SCF is consistent with that long term trend (although one has to also look at the other surveys and tax data to fully establish this trend). In short, neoliberalism has increasingly narrowed and concentrated the base of those who benefit from it. Of course if this is true it will have many implications for society, politically, and for future macroeconomic dynamics.
Hope this helps with the question. Paul